GAO

advertisement
_ ..“- ,.... -..
Unit&
GAO
Stqtes
General
Accounting
Office
Report to the Chairman, Legislation
‘axd Fktimal Security Subcommittee,
Cgmx&tee on Govement Operations,
House af Representatives
Stable Design and
Adequate Testing Must
Precede Decision to
Deploy
GAO/IMTEC-SO-61
I
.
‘::
.I ‘::
. .
.
.. ..
:.
.
I
,.
.
. .
.‘:‘.
.:
:
.’
.‘i
.
. .
: : ;
‘.
..
i.
‘.
; ’
. .
‘1,
.,
..
.’
.
’
. ..
.
..
..
..
.
’
GAO
United States
General Accounting
Office
Washington, D.C. 20548
Information
Technology
Management
Division
and
B-239476
July 6, 1990
The Honorable John Conyers, Jr.
Chairman, Legislation and
National Security Subcommittee
Committee on Government Operations
House of Representatives
Dear Mr. Chairman:
This report discusses why the Strategic Defense Initiative Organization will not be able to
support currently scheduled full-scale development or deployment decisions on any part of
Phase I of the Strategic Defense System. The organization will not be able to support these
decisions until it has solidified the system architecture, conducted integrated system-level
testing in real time involving actual pieces of hardware and software, and until the
Department of Defense adheres to technical and programmatic oversight, commensurate
with such a system.
As arranged with your office, unless you publicly announce the contents earlier, we plan no
further distribution of this report until 30 days from the date of this letter. We will then send
copies to other appropriate congressional committees; the Director, Strategic Defense
Initiative Organization; the Director, Office of Management and Budget; the Secretary of
Defense; and other interested parties.
This report was prepared under the direction of Samuel W. Bowlin, Director for Defense and
Security Information Systems, who may be reached at (202) 275-4649. Other major
contributors are listed in appendix IV.
Sincerely yours,
Ralph V. Carlone
Assistant Comptroller General
Executive Summary
Purpose
to develop and deploy the first of a multiple phase Strategic Defense
System designed to shield the United States from a massive Soviet
nuclear ballistic missile attack. At the request of the Chairman, Legislation and National Security Subcommittee, House Committee on Government Operations, GAO reviewed the Strategic Defense Initiative
Organization’s (SDIo's) test and evaluation plans. These plans are to support a 1993 executive decision by the President on deploying Phase I
and a 1994-1995 acquisition decision by the Department of Defense to
enter full-scale development. Specifically, GAO was asked to determine
l
what test and evaluation requirements were established for making
development and deployment decisions for the first phase, and
. whether SDIO'S approach will meet those requirements and support a
decision on deployment.
Background
The Strategic Defense System is being designed to destroy thousands of
incoming ballistic missiles and warheads. This will require detecting the
missiles and warheads, discriminating them from hundreds of thousands
of decoys, precisely tracking the missiles and warheads, and then
destroying them-all within about 35 minutes or less, and despite the
enemy’s concerted effort to defeat the system. The Strategic Defense
System is, by far, the most complex, technologically challenging system
ever attempted. As currently envisioned, a sophisticated communications system will link together hundreds, or possibly thousands, of computer-operated components in space and on the ground. The system will
consist of several subsystems (elements), which include surveillance
satellites, space- and ground-based weapons, and ground-based subsystems to command and control the satellites and weapons.
The Strategic Defense System is being developed in several phases.
Phase I will be built using existing technologies; later phases would use
more exotic technologies, such as lasers and particle beams. On September 17, 1987, the Secretary of Defense directed that Phase I enter
the concept demonstration and validation stage. This stage is important
because enough information must be developed to show that the system
is feasible before a decision is made to enter full-scale development.
Because Phase I cannot be demonstrated outside of an actual ballistic
missile attack, much of this information can be generated only through
computer simulation and modeling.
Page 2
GAO/IMTEG9OH)Il
SD1 Needs Stable Design and Adequate
Testing
Executive
Summary
In 1988, SD10 formally adopted a design (architecture) for Phase I. However, in January 1990 SD10 changed the design dramatically by incorporating a new space-based weapon, called Brilliant Pebbles. The idea
behind Brilliant Pebbles is that thousands of small, relatively autonomous weapons would be deployed to intercept Soviet intercontinental
ballistic missiles.
Results in Brief
The President is scheduled to make a decision in 1993 on deploying
Phase I. His options include deploying, delaying, or cancelling Phase I.
Public Law 99-145 states that a Strategic Defense System may not be
deployed in whole or in part unless the President certifies to the Congress that the system can fulfill its mission. SD10 may be able to justify a
decision to delay or cancel Phase I; however, SDIO will not be able to
support an informed decision to deploy it. To do so requires a stable
design, sufficient testing and evaluation, and according to SDIO, minimum
funding levels. However, it is highly unlikely any of these conditions
will be met.
The addition of Brilliant Pebbles in January 1990 has reduced, changed,
or eliminated the need for some of the space-based elements in the 1988
design. At this point, it is uncertain exactly what Phase I will consist of
in terms of elements or what functions they will perform. Additionally,
while SDIO initially required real-time integrated system-level testing
prior to a presidential deployment decision and a full-scale development
decision, SD10 officials now state that such detailed tests will not be conducted prior to either decision. Furthermore, according to SDIO, because
of fiscal year 1990 funding cuts, the full-scale development decision has
been delayed until 1994-1995, and research efforts have been scaled
back, further reducing the amount of information that will be available
for a presidential deployment decision. SD10 officials believe they will be
able to support a presidential deployment decision with less information
than was originally desired, but state that this will be done at increased
risk.
Phase I has not received the scrutiny and oversight that Defense initially envisioned. The Defense Acquisition Board did not conduct its
scheduled 1989 review, nor has it assessed how Brilliant Pebbles affects
the design and test and evaluation requirements. The lack of effective
agency oversight has contributed to the failure of other automated
weapons systems, none of which matches the scale and technological
complexity of Phase I. Consequently, any executive decision in 1993 to
deploy Phase I would be premature and fraught with high risk.
Page 3
GAO/IMTEG90-61
SD1 Needs Stable Design and Adequate
Testing
Executive
Summary
Principal Findings
SD1Design and Test Plans
in State of Flux
Several problems are hampering SDIO’S design and test plans. First, Brilliant Pebbles radically changed the design and sent the program into a
state of flux. SDIO has not yet solidified the role of Brilliant Pebbles or
what elements will be in the final design. This causes problems because
without a stable design detailed Strategic Defense System requirements
cannot be determined. Setting requirements is especially important for
Phase I where weapon, sensor, and communication systems are interdependent. Changing a requirement for a weapon system could significantly affect sensor and communications systems as well. Further, an
unstable design increases the probability that system requirements will
not be adequately determined and sufficient testing will not be performed to ensure that the system works. Nevertheless, Defense has
requested $265 million for full-scale development of one of the sensor
elements, the Boost Surveillance and Tracking System, for fiscal year
1991 even though Brilliant Pebbles may eliminate or reduce the requirements for this system.
Second, SD10 does not plan to conduct integrated system-level testing by
the scheduled 1993 presidential decision. SDro-funded studies and test
plans have cited the importance of running real-time integrated systemlevel tests, which combine computer simulation and actual hardware
and software prototypes with human intervention. Furthermore, most
of the system tests and evaluations so far have been based on the preBrilliant Pebbles design. Consequently, much of the testing and analyses
may no longer be relevant and will have to be repeated.
Also, according to SDIO, its ability to support an informed deployment
decision depended on a minimum funding level of $4.6 billion for fiscal
year 1990. However, funding was cut by 20 percent According to SDIO,
such a cut will reduce the amount of research performed, delay SDIO’S
full-scale development decision by 2 years, and reduce the amount of
information available t,o the President. Nevertheless, SDIO officials have
stated that they will be able to support a presidential decision on the
program even though less information will be available.
Page 4
GAO/IMTEGSOBl
SD1 Needs Stable Design and Adequate
Testing
Executive
Summary
Original Program
Oversight Reduced
Because of the complexity, uniqueness, and cost of Phase I, the Defense
Acquisition Board initially was to follow a highly structured oversight
system of annual reviews looking at cost growth, changing requirements, and test and evaluation results. However, the October 1989 program review scheduled by the Defense Acquisition Board did not occur.
According to Defense officials, the review was cancelled because the
fiscal year 1990 budget had not yet been finalized and because Defense
was reassessing the amount of oversight needed for the Strategic
Defense System. Even though a Defense Acquisition Board review is
scheduled for June 15, 1990, major changes to the Phase I design and
significant reductions to test and evaluation requirements have already
occurred without the Board’s review and evaluation.
the Secretary of
Defense
has stabilized the design and has demonstrated the effectiveness of the
system through integrated system-level tests in real time, using system
hardware and software prototypes with human intervention. Further,
the Secretary should ensure that required oversight by the Defense
Acquisition Board be followed. The Board should more closely monitor
system design, development, testing, and evaluation. GAO is making
other recommendations in chapter 4.
recommends that the Congress not fund full-scale development for
any element of Phase I, including $266 million for the Soost Surveillance
and Tracking System in fiscal year 1991, until SDIO has stabilized the
design and has demonstrated the effectiveness of the system through
integrated system-level tests in real time, using system hardware and
software prototypes with human intervention.
Recommendation to
the Congress
GAO
Agency Comments
As requested by the Chairman’s office, GAO did not obtain official
agency comments on a draft of this report. However, GAO discussed the
information contained in this report with SD10 officials and has incorporated their comments where appropriate.
Page 5
GAO/IlHTEC908I
SD1 Needs Stable Design and Adequate
Testing
contents
Executive Summary
Chapter 1
Introduction
Chapter 2
Unstable Architecture
and Test Plans Will
Not Support a
Presidential
Deployment Decision
in 1993
Chapter 3
Program Management
Oversight Has Been
Reduced
Chapter 4
Conclusions and
Recommendations
Appendixes
2
Genesis/Evolution of the SD1 Program
Objective of a Strategic Defense System
Description of Phase I
Importance of Concept Demonstration and Validation and
Defense Acquisition Oversight
Constraints Affecting SDIO’s Approach to Concept
Demonstration and Validation
Objectives, Scope, and Methodology
Brilliant Pebbles Sends Phase I and Strategic Defense
System Into State of Flux
SD10 Will Not Perform Integrated System-Level Testing
Prior to 1993
SD10 Asserts Funding Cuts Will Reduce Information
Needed for an Informed Decision in 1993
Oversight for Phase I More Stringent Than Typical
Acquisition
1989 Annual DAB Review Did Not Occur; Decisions Have
Been Made Without DAB Review and Approval
Congress Cuts Funding for Operational Testing and
Evaluation Until Architecture Is Solidified
8
8
9
11
13
14
15
18
18
22
24
25
25
26
27
29
Recommendations to the Secretary of Defense
Recommendation to the Congress
31
31
Appendix I: Department of Defense Major Systems
Acquisition Stages
Appendix II: Strategic Defense System Elements of
Phase I
32
Page 6
GAO/IMTEGBOSl
SD1 Needa Stable Design and Adequate
34
Testing
Contents
Appendix III: Test Beds Supporting Demonstration and
Validation of the Strategic Defense System
Appendix IV: Major Contributors to This Report
Figures
Figure 1.1: Phases of a Ballistic Missile Attack
Figure 1.2: Phase I Architecture Incorporating Brilliant
Pebbles
35
36
10
12
Abbreviations
DAB
GAO
IMTEC
SD1
SD10
Page 7
Defense Acquisition Board
General Accounting Office
Information Management And Technology Division
Strategic Defense Initiative
Strategic Defense Initiative Organization
GAO/IMTEZGgoBl
SD1 Neede Stable Dee&n and Adequate
Testing
Chapter 1
Introduction
The Strategic Defense Initiative (SDI) is a program to develop and deploy
a Strategic Defense System to shield the United States against a Soviet
nuclear ballistic missile attack. Research and development for the SDI
program began in 1984, and in 1987 a Strategic Defense System was
approved for acquisition. This system is to be developed in multiple
phases. Phase I, the subject of this report, is in the concept demonstration and validation stage of the Department of Defense’s acquisition
process.’
The Strategic Defense System will be an expensive undertaking. The
Strategic Defense Initiative Organization (SDIO) cost estimates range
from $69.6 to $87 billion for completion of Phase I development and
deployment-with
an additional $10.1 billion for Phase I operation and
support until deployment is completed* While the cost of follow-on
phases is unknown, SDIO estimates that, in addition to the Phase I costs,
approximately $26.4 billion will be needed for fiscal years 1990 through
1994 to pursue non-Phase 1 research and development. GAO has recently
reported to the Congress that the above figures are, at best, optimistic.’
Genesis/Evolution of
the SD1 Program
On March 23, 1983, President Reagan called for a comprehensive scientific research effort to develop a system that would render nuclear ballistic missiles impotent and obsolete. In January 1984, SD1 was
established as a research and technology development program, and in
April 1984, Defense formally chartered the Strategic Defense Initiative
Organization as the agency responsible for managing Defense’s efforts.
After several years of research, SDIO decided in 1987 to develop and
deploy the system in phases. SDIO felt that working in phases would
allow SDIO to be prepared for an early deployment, if such a decision
were made, and to respond to changing threats. Phase I of the system is
to be based on currently available technologies, while later phases are to
incorporate technologies that are expected to be available in the future,
such as neutral particle beams and lasers. Phase I is not intended to be a
“total defense”, but is being designed to destroy a certain percentage of
some intercontinental and sea-launched ballistic missiles (the exact percentages are classified). Creating a full strategic defense capability will
‘The Department of Defense’s mdor system acquisition process is supposed to provide a single
approach to designing, developing, implementing, and maintaining major weapons systems. (The five
stages of the acquisition process are discussed in appendix 1.)
Page 9
GAO/MTFE-SO-61
SD1 Needs Stable Design and Adequate
Testing
Chapter 1
Introduction
require the deployment of the follow-on phases. Therefore, a critical
aspect of the first phase is the capability to evolve and support future
phases. Further, the system is not being designed to destroy sea-and airlaunched cruise missiles.:’ Other costly systems independent of the SDI
program will be needed to defend against these weapons.
From the beginning, SDIOhas had one overall goal-to conduct a vigorous research and technology program that would provide the basis for
an informed decision regarding the deployment of a Strategic Defense
System. The current Phase I program is intended to support an executive decision on deployment by the President in 1993 and an acquisition
decision on full-scale development by Defense in 1994 or 1995.
According to SDIO test and evaluation officials, the President will have a
range of options including deploying, delaying, or cancelling Phase I.
However, if the President decides to deploy the system, Phase I development will not be consistent with Defense’s prudent acquisition policies
specifying that deployment decisions be made after full-scale
development.
Objective of a
Strategic Defense
System
The Strategic Defense System is based on a layered defense concept; the
system is supposed to intercept a missile or a deployed warhead, as it
travels toward its target. The basic concept is that separate sensor,
weapon, and command and control systems (SDIO refers to these separate systems as elements) would be in space and on the ground. The
system would be tied together by a complex communications network
and a sophisticated computer-based command and control element.
During an attack, the system would have to function in an extremely
hostile environment, including nuclear explosions and a concerted effort
by the enemy to destroy the Strategic Defense System.
The threat the system would be facing, along with the environment in
which it would be operating, creates a unique and demanding challenge.
First, the space-based elements of the Strategic Defense System would
have to be able to detect and begin tracking thousands of missiles almost
immediately after launch, and then intercept and destroy some of them,
Those missiles that are not destroyed would release warheads, along
with decoys intended to confuse the system, forcing the system to track
hundreds of thousands of objects. Also at this point, space-based elements of the system would have to distinguish the warheads from the
%kuise missiles are guided missiles that have terrain-seeking radar and fly at moderate speed in low
altitude,
Page 9
GAO/IMTEG904il
SD1 Needs Stable Design and Adequate
Testing
Chapter 1
Introduction
decoys, and intercept and destroy some of the warheads. The system
would need to continue discriminating and tracking the surviving warheads so that the ground-based elements could intercept more warheads
before they hit the United States. These functions would have to be
tightly coordinated and performed in less than 35 minutes-all this with
nuclear warheads exploding, anti-satellite weapons attacking the
system, and the enemy trying to disrupt communications and computer
operations. Figure 1.1 describes the phases of a ballistic missile attack.
Figure 1.1: Phases of a Ballistic Missile Attack
Warheads and
Decoys
0
V
QVOL
0p 8
e@V
Midcourse
About 20 minutes
CD
Terminal
30 to 60
seconds
Post-Boost
IO’s of seconds
to 10’s of minutes
@a
7
Source. Adapted from SDI Technology Survivability
May 1988
Page 10
GAO/KMTEG9041
Boost
Several 1O’s
to loo’sof
seconds
and Software, Offlce of Technology Assessment.
SD1 Needs Stable Design and Adequate
Testing
Chapter 1
Introduction
Description of Phase I
In 1988, a Phase I design, made up of seven elements, was approved by
the Defense Acquisition Board. The design consisted of two space-based
sensors (Boost Surveillance and Tracking System and Space Surveillance
and Tracking System); a space-based weapon (Space-Based Interceptor);
two ground-based sensors (Ground Surveillance and Tracking System
and Ground-Based Radar); and a ground-based weapon (Ground-Based
Interceptor), all managed by a command and control system (Command
Center Element). (Appendix II describes each element.) The six surveillance and weapon elements would be highly interdependent and rely
heavily on the command and control element, along with a complex communications system. The individual elements would work together as an
integrated system to detect, track, discriminate, and destroy ballistic
missiles and their warheads. At the heart of the system is a large, distributed, real-time computer software system which, by some estimates,
could have 40-100 million lines of code. The complexity and risk of this
undertaking cannot be overstated. We have reported repeatedly on
Defense systems that were far less complex, contained far fewer lines of
code, but have yet to perform as intended.’
In January 1990, SDIO decided to include a new weapon concept, called
Brilliant Pebbles, in Phase I. Brilliant Pebbles involves several thousand
individual interceptors orbiting the earth in order to detect and destroy
a target by smashing into it at high speed. Brilliant Pebbles is supposed
to improve system survivability and reduce costs by dispersing
thousands of space-based interceptors that are smaller and more autonomous than the Space-Based Interceptor. The Space-Based Interceptor
would house a number of interceptors, but unlike Brilliant Pebbles, it
would rely on other satellites for tracking, targeting, and communications. By making interceptors autonomous, they would no longer need to
rely on other satellites to perform these functions. Further, because each
interceptor would work independently, the system’s survivability would
be increased, in principle, because the loss of any one interceptor would
not greatly affect the system’s overall effectiveness. Figure 1.2 shows
Phase I with the six elements and Brilliant Pebbles.
Page 11
GAO/IMTEG9O-61
SD1 Needs Stable Desigu and Adequate
Testing
Chapter 1
Introduction
Fiaure 1.2: Phase I Architecture Incorporating Brilliant Pebbles
BSTS is Boost Surveillance
and Tracking System
GBI is Ground-Based
Interceptor
GBR is Ground-Based
Radar
GSTS is Ground $urveillance
and Tracking System
SSTS is Space Surveillance
Tracking System
RVs are Reentry Vehicles (nuclear warheads)
funded three Brilliant Pebbles research and development studies
during 1939 and 1990. Studies completed by the Defense Science Board
and the JASONs-a group of scientists who periodically provide teehnical support to the Department of Defense-determined
that the Brilliant Pebbles concept was technically feasible. Along with these two
studies, Defense issued a Space-Based Architecture Study which
reviewed the four space-based elements of Phase I-Space Surveillance
and Tracking System, Boost Surveillance and Tracking System, the
Space-Based Interceptor, and Brilliant Pebbles-and defined and justified a recommended space-based architecture. The Space-Based Architecture Study recommended that research continue on both Brilliant
SD10
Page 12
GAO/IMTEC-90-61
SD1 Needs Stable Design and Adequate
Testing
Chapter 1
Introduction
Pebbles and the Space-Based Interceptor and that modified versions of
the Boost Surveillance and Tracking System and the Space Surveillance
and Tracking System sensors be included.
has decided to eliminate the Space-Based Interceptor from Phase I,
claiming that Brilliant Pebbles will cut costs by 20 percent, that is, from
$87 to $69.6 billion. However, it is unclear whether the costs for all
Phase I requirements have been incorporated into the $69.6 billion, Further, neither figure includes the costs for operating and maintaining the
system.
SDIO
Importance of Concept
Demonstration and
Validation and
Defense Acquisition
Oversight
On September 17, 1987, the Secretary of Defense directed that Phase I
of the Strategic Defense System enter the concept demonstration and
validation stage of Defense’s major systems acquisition process. SDIO’S
basic acquisition strategy is to complete Phase I while researching later
phases of a Strategic Defense System. After successfully completing concept demonstration and validation, Phase I could enter full-scale
development.
According to the Defense Systems Management College, concept demonstration and validation may be the most critical of all acquisition stages.
The primary purpose of this stage is to validate the feasibility of Strategic Defense System concepts and technologies, and show that the risk
is low enough to enter full-scale development. According to the College,
the Secretary of Defense’s decision to place a major system in full-scale
development is extremely important. During full-scale development, the
system is built, tested, and ready for full-rate production. Not only will
development consume enormous resources, but major systems, at this
point, take on a life of their own and are seldom cancelled.
To help Defense gain sufficient design and development information,
major system acquisitions are reviewed by the Defense Acquisition
Board (DAB). The purpose of the DAB, which is chaired by the Under Secretary of Defense for Acquisition, is to oversee major Defense acquisitions, The DAB reviews each acquisition stage to ensure that every
program is ready to proceed into more advanced stages of development
or production. Each stage of the acquisition process, as well as each
milestone decision, is to be supported by test and evaluation. The purpose of test and evaluation is to help ensure the timely development,
production, and fielding of systems that meet users’ requirements and
Page 13
GAO/IMTEC-9041
SD1 Needs Stable Deslgu and Adequate
Testing
Chapter 1
Introduction
perform as intended. The DABrelies on information from test and evaluation to determine whether a system is ready to enter full-scale development. The DAB reviews also ensure that plans for later stages follow
sound acquisition management practices.
The complexity, cost, and uniqueness of the Strategic Defense System
has prompted the need for a more detailed program review process. For
example, the DAB developed a guidance document called the Implementor, which provides a framework for additional oversight. Although
formal DAB reviews are usually only required at major milestones, the
Implementor recommends yearly DAB reviews. Further, SDIO must submit
additional documentation to the DAB to ensure program goals are being
met.
Constraints Affecting
SDIO’s Approach to
Concept
Demonstration and
Validation
Constraints exist which affect how the Strategic Defense System can be
demonstrated. Although the 1972 Anti-Ballistic Missile Treaty permits
research, it limits the development, testing, and deployment of different
types of strategic defense systems and components. According to SDIO'S
1989 report to the Congress on SDI compliance with the treaty, research
includes conceptual design and laboratory testing. Development occurs
after research but precedes full-scale testing of systems and components
designed for actual deployment. The development, testing, and deployment of launchers, interceptors, and radars are restricted. Also, the
extent to which the system can demonstrate its ability to interact in a
nuclear environment is constrained by limits on nuclear testing.
Because the Strategic Defense System cannot be tested in its operational
environment, many system capabilities must be demonstrated through
computer modeling and simulation. For example, in simulations,
software models would mimic the behavior of sensors and weapons and
be used in place of the actual elements to evaluate system performance.
A ballistic missile attack from launch to impact must also be simulated
in software to prompt the element models to respond. Eventually, some
of these tests would involve actual prototypes5 of weapons and sensor
hardware and software interacting in the simulation. This type of integrated system-level testing would be used to evaluate the performance
of elements within the context of the entire Strategic Defense System.
Thus, the ability to simulate the interaction of the systems’ hundreds of
“A prototype is an original or model on which a later item is formed or based. A prototype is usually
built during the concept demonstration and validation stage and tested prior to the full-scale development decision.
Page 14
GAO/lMTEX-9061
SD1 Needs Stable Design
and
Adequate
Testing
Chapter 1
Introduction
computers with their many millions of lines of software code becomes of
paramount importance in demonstrating whether the system can perform its mission.
Early in the SDI program, SD10 recognized the need for integrated systemlevel testing prior to a full-scale development decision. Defense policies
state that models or simulations can support a full-scale development
decision and that prototyping is one of the most powerful tools available
for determining system feasibility and capability. Consequently, SDIO initially planned to build a simulation framework to conduct tests using
software and hardware prototypes. SD10 studies and contractor documents have supported the need for such test capabilities before a fullscale development decision. However, because SDIO believed building
such a simulation framework would be very complex, it explored other
approaches for demonstrating and validating the Phase I system.
decided instead to use multiple test beds.l; Under this approach,
each test bed would address critical strategic defense issues such as
system performance, command and control, communications, and
sensing and tracking. The system performance test bed, known as the
System Simulator, would be at the core of this test environment. The
System Simulator, using computer models of individual elements, would
perform end-to-end, system-level tests in which all aspects of Phase I’s
performance would be evaluated while under a simulated ballistic missile attack. Although the System Simulator, or some other simulation
framework, might evolve to include real-time integrated hardware-,
software-, and human-in-the-loop capabilities, SDIO has no plans to
accomplish this prior to either the presidential decision on deployment
or the full-scale development decision.
SDIO
Objectives, Scope, and
Methodology
On June 5, 1989, the Chairman, Legislation and National Security Subcommittee, House Committee on Government Operations, requested that
we assess the Strategic Defense Initiative Organization’s program activities, test and evaluation requirements, and decisions on the development
and deployment of Phase I of the Strategic Defense System. Specifically,
the objectives of our review were to determine (1) what test and evaluation requirements were established for making development and deployment decisions for the first phase of the Strategic Defense System and
“Test beds are sites, facilities, or activities used for testing models or prototypes. SDIO refers to its
network of geographically dispersed test facilities as the National Test Bed.
Page 15
GAO/IMTEGSOBI
SD1 Needs Stable Design and Adequate
Testing
Chapter 1
Introduction
l
(2) whether the current program approach will meet those requirements
and support a decision on whether to deploy.
To address our objectives, we interviewed a wide range of officials
involved in the Strategic Defense Initiative. SDIO representatives
included: (1) the Phase I System Engineer, as well as other Engineering
Office officials, (2) test and evaluation officials, (3) Brilliant Pebbles
Program Office officials, and (4) National Test Bed Program Office officials. We discussed technical issues with members of the Phase One
Engineering Team, which acts as a consultant to SDIO. Team members are
experts in ballistic missile defense and include physicists, software and
electrical engineers, and simulation analysts. We also met with officials
from Martin Marietta, the prime contractor for the National Test Bed,
and officials from General Electric Company, the system engineer and
integration contractor for the Strategic Defense System. General Electric
has the responsibility for defining Phase I demonstration and validation
requirements, including those for test and evaluation. We met wjth
Army and Air Force officials involved in major SD1 research and development as well as Defense Acquisition Board officials. Additionally, we
consulted with ballistic missile defense experts at Sandia and Los
Alamos National Laboratories.
Our approach was to obtain a thorough understanding of the underlying
engineering concepts for the Strategic Defense System by (1) obtaining
and analyzing key system design and requirements documents, (2) interviewing and consulting with experts from the ballistic missile defense
community, and (3) visiting and assessing research and development
sites. We reviewed pertinent management, technical, and contract documents provided by SDIO, SDIO contractors, the Air Force, the Army, Los
Alamos National Laboratory, and the Defense Science Board. We analyzed and compared system documents such as the 1988 System
Description Document, the 1990 System Description Document, the Brilliant Pebbles System Description Document, and the Space-Based Architecture Study. Furthermore, we analyzed and compared test and
evaluation documents including the 1987 and 1989 Test and Evaluation
Master Plan, the Phase I Integrated Test Plan, the System Simulator
requirements document, the technical specifications and software for
the current System Simulator, the Brilliant Pebbles Integrated Test Plan,
and the Brilliant Pebbles Test and Evaluation Master Plan. Through our
analyses we were able to identify the chronology of events that led to
SDIO's current test and evaluation approach in support of decisions on
deployment and full-scale development. We did not, however, validate
test and evaluation results.
Page 16
GAO/IMTEC-90-61
SD1 Needs Stable Design and Adequate
Testing
Chapter 1
Introduction
Our work was conducted at SD10 Headquarters, Washington, D.C.; the
National Test Bed Joint Program Office, Falcon Air Force Base, Colorado
Springs, Colorado; Electronic Systems Division, Boston, Massachusetts;
the Army Strategic Defense Command, Huntsville, Alabama; Sandia
National Laboratory, Albuquerque, New Mexico; Los Alamos National
Laboratory, Los Alamos, New Mexico; the Phase One Engineering Office,
Crystal City, Virginia; the Defense Acquisition Board, Washington, D.C.;
General Electric Corporation, Blue Bell, Pennsylvania; and Martin Marietta Corporation, Colorado Springs, Colorado.
We discussed the facts in this report with SDIO officials and have incorporated their comments where appropriate. However, in accordance
with the requester’s wishes, we did not obtain official agency comments
on a draft of the report. We conducted our work from June 1989
through May 1990, in accordance with generally accepted government
auditing standards.
Page 17
GAO/IMTEC90-61
SD1 Needs Stable Design and Adequate
Testing
Unstable Architecture and Test Plans Will Not
Support a Presidential Deployment Decision
in 1993
The President is scheduled to make a decision in 1993 on deploying
Phase I of the Strategic Defense System.’ Public Law 99-145 states that
a Strategic Defense System may not be deployed in whole or in part
unless the President certifies to the Congress that the system can survive, meaning that it can maintain enough effectiveness during a war to
fulfill its mission. However, by 1993 SD10 will not have conducted integrated system-level tests designed to demonstrate that the entire Phase I
system will work as planned, so it will not be able to support an
informed deployment decision by the President.
The recent inclusion of Brilliant Pebbles has fundamentally changed the
1988 Phase I architecture by potentially reducing or eliminating the
need for two sensors and one weapon component of the system. This
change has significantly destabilized the architecture, resulting in a dramatic restructuring of Phase I. According to SDIO test and evaluation
officials, the new Phase I architecture will not be solidified until 1991
thus decreasing the level of system testing that can be performed by
1993. Furthermore, according to SDIO’S Director, because of fiscal year
1990 funding cuts, research will be scaled back, reducing the information available on which to make a deployment decision in 1993.
Brilliant Pebbles Sends
Phase I and Strategic
Defense System Into
State of Flux
In January 1990, Brilliant Pebbles was formally incorporated into the
Phase I design. However, exactly what role Brilliant Pebbles will play in
the Phase I architecture and even what pieces of Phase I will be
deployed is uncertain and consequently, has put the design of the Strategic Defense System into a state of flux. The result is a destabilized
architecture and a major restructuring of SDIO’S program.
Stable Integrated
Architecture Needed for
System Development
The SDIO Phase I System Engineer has stated that his first goal in support of the presidential deployment decision is to create “a complete and
coherent system design with definition of all elements and inter-element
interfaces.“’ It is important in systems development to have a stable,
integrated architecture before detailed system requirements can be
determined. A stable architecture provides the blueprint for design and
development of each element. An unstable architecture causes confusion
and increases the probability that system requirements and integrated
‘SD10 is conducting the SD1program to support a decision by the President prior to the completion of
his current term. Much of the specific program guidance and direction, as well as the timing of the
presidential decision is classified.
“Interfaces are the internal and external communication paths within and outside of the system.
Page 18
GAO/IMTEC-90-61
SD1 Needs Stable Design and Adequate
Testing
Chapter 2
Unstable Architecture
and Test Plans Will
Not Support a Presidential
Deployment
Decision
in 1993
test objectives will not be met, and that the system will not perform as
intended.
The Strategic Defense System is a “system-of-systems” and accordingly,
the elements are all interdependent. A requirement change for one element may affect all the others. For example, in the 1988 Phase I architecture, the Boost Surveillance and Tracking System and the Space
Surveillance and Tracking System had a requirement to combine their
tracking data to provide targeting information for the Space-Based
Interceptor. This combined data would also be transmitted to the
ground-based elements to help them identify targets. The interfaces
must be precisely defined to support such data sharing. If the Space Surveillance and Tracking System’s mission changes, then all the elements
must reflect the change as well. Accordingly, to ensure integrated
system development, each element developer must be aware of all the
other elements and the respective interfaces when designing each
system.
Brilliant Pebbles
Destabilizes Phase I
Architecture
Major program restructuring is going on to incorporate Brilliant Pebbles
and stabilize the architecture. Brilliant Pebbles is now the focal point of
SDIO’S
efforts. Its effect on the Phase I architecture is far-reaching.
As discussed in chapter 1, the 1988 Phase I architecture was made up of
seven highly integrated elements. Command and control relied on rapid
communication and data sharing among elements. For example, during a
battle, the Boost Surveillance and Tracking System would begin tracking
boosters and would be one of several systems to alert national decisionmakers. The Boost Surveillance and Tracking System would then
transfer the information to the Space Surveillance and Tracking System,
which would begin tracking and pass intercept information to the SpaceBased Interceptor. At the same time, battle information would be sent to
the ground-based weapons and sensors. The Ground Surveillance and
Tracking System and the Ground-Based Radar would continue tracking
and the Ground-Based Interceptor would destroy the warheads before
they reenter the earth’s atmosphere. Throughout the battle, command
and control information would be centralized at the ground-based Command Center Element.
The addition of Brilliant Pebbles fundamentally changed this highlyintegrated approach to strategic defense. Brilliant Pebbles provides for
several thousand self-contained interceptors that can detect and destroy
missiles independently of the other Phase I elements. As a consequence,
Page 19
GA0/1MTEG90-61
SD1 Needs Stable Design and Adequate
Testing
Chapter 2
Unstable Architecture
and Test Plans Will
Not Support a Presidential
Deployment
Dfxision in 1993
interfaces, sensing and tracking capabilities, communications, and the
numbers of elements needed-both in terms of types and constellations
(numbers of satellites)-will
have to be reassessed. For example, the
Boost Surveillance and Tracking System may not be required as a Phase
I element, and Brilliant Pebbles will add thousands of independent
interceptors to the system. Furthermore, Brilliant Pebbles will eliminate
the need for the Space-Based Interceptor, could reduce tracking requirements placed on the Space Surveillance and Tracking System, and could
expand the mission requirements for the Ground Surveillance and
Tracking System. Unlike the Space-Based Interceptor, whose battle management functions were highly dependent on ground-based command
and control instructions, Brilliant Pebbles, after being turned on by the
commander, could be autonomous.
Phase I Architecture
Remains Uncertain
Defense officials have stated that the elements that will make up the
Phase I architecture have not yet been solidified. However, several scenarios have been given. An Office of the Secretary of Defense official
stated that a space-based system made up of the Boost Surveillance and
Tracking System and BrilIiant Pebbles could be initially deployed as the
Phase I architecture, follo;ved by the Phase I ground-based elements.
Another scenario is that the Boost Surveillance and Tracking System
and Brilliant Pebbles would be deployed with a terminal interceptor!
(the High Endoatmospheric Interceptor), which is not currently included
in the Phase I design.
In April 1990, the Director of SDIO told the Congress that because of Brilliant Pebbles potential for independent operation, the Phase I architecture may not require the Boost Surveillance and Tracking System and
the Space Surveillance and Tracking System or may require less capability from them. However, in May 1990, SDIO officials stated that the
Phase I architecture will include Brilliant Pebbles and all the 1988 Phase
I architecture elements, with the exception of the Space-Based Interceptor. The differences among such scenarios and the uncertain architecture point to the instability of basic concepts underlying the Strategic
Defense System.
Unless the architecture is stabilized and the respective elements are
developed, designed, and tested as an integrated system, SD10 will not
“Terminal interceptors are intended to destroy warheads after they have entered the earth’s
atmosphere.
Page 20
GAO/lMTEC-9041
SD1 Needs Stable Design and Adequate
Twting
Chapter 2
Unstable Architecture
and Test Plans Will
Not Support a Presidential
Deployment
Decision in 1993
know whether the system will work as planned. The process of interconnecting the Phase I elements so that together they can effectively
accomplish the strategic defense mission is called system integration.
Successful system integration should result in the design and delivery of
complete elements that will work in concert with the other Phase I elements. Two of the elements, however, are preceding the rest of the
system. Because of presidential guidance, SDIOhas expedited the
research and development of Brilliant Pebbles. Consequently, Brilliant
Pebbles has been placed in an accelerated acquisition program. The
Boost Surveillance and Tracking System, which may or may not be
needed if Brilliant Pebbles is successful, is approaching a full-scale
development decision. Defense has requested $265 million for full-scale
development of the Boost Surveillance and Tracking System in fiscal
year 1991. The Boost Surveillance and Tracking System is a multi-missioned satellite that supports other Defense programs. Defense documents show that SDIOwill provide $1.3 billion (or 99 percent) of the
system’s funding through fiscal year 1991. The Boost Surveillance and
Tracking System’s specific functions for the Strategic Defense System
are to detect missile launches, acquire and track boosters, and assess the
number of boosters and warheads that have been destroyed. If Brilliant
Pebbles and the Boost Surveillance and Tracking System are allowed to
advance before the requirements of the other elements are clearly
defined, significant integration problems and costs could emerge.
Phase I System Test Plans
and Results May Be Moot
test and evaluation officials claim that including Brilliant Pebbles in
Phase I has destabilized the architecture and that a firm design will not
be available before 1991. For example, they have stated that a Preliminary Systems Requirements Review that was scheduled for May 1990
will be deferred for at least a year. Furthermore, SDIO test and evaluation officials are rewriting test plans to incorporate Brilliant Pebbles.
Because most of the test and evaluation efforts have been based on the
1988 Phase I architecture, much of the data and analysis may no longer
be relevant. For example, if, as mentioned above, the Space Surveillance
and Tracking System’s mission is changed and it no longer provides
tracking information, then all of the modeling and simulation done so far
is useless because all the interfaces and data paths have changed. Any
change to one element causes this rippling effect across the system.
SD10
Further, system-level tests and the resulting requirements
included Brilliant Pebbles. Accordingly, system tests must
and actual simulations rerun to include them. Because the
elements have not included Brilliant Pebbles in their tests,
Page 21
GAO/IMTEC-90-61
have not
be rewritten
other Phase I
analyses, and
SD1 Needs Stable Design and Adequate
Testing
Chapter 2
Unstable Architecture
and Test Plans Will
Not Support a Presidential
Deployment
Decision in 1993
development, element test plans and simulations must also be
reassessed.
SD10 Will Not Perform
Integrated SystemLevel Testing Prior to
1993
snlo-funded studies and test plans cite the importance of conducting
integrated system-level tests in real time, using actual system hardware
and software with human intervention. For example, the Advanced Simulation Framework Study recommended that component hardware and
software models at various geographic locations be tested and evaluated
to see how they work in real-time operations. Moreover, SDIO'S Integrated Test Plan notes the importance of such tests before making a
decision on full-scale development. According to the plan, integrated
system testing minimizes costs by integrating and building upon existing
experiment assets, data, and results, and sharing a common test bed.
However, an integrated system-level test in real time using hardware
and software and a human-in-the-loop will not be demonstrated prior to
the planned 1993 decision on deployment. SDIO officials feel they will be
able to support a presidential deployment decision with less information
than was originally desired, but state that this will be done at increased
risk.
Value of Integrated
System-Level Testing
Integrated system-level testing will help confirm whether the individual
elements and the strategic defense commander can successfully interact
in real time. Actual hardware and software in the test environment will
provide significantly more information about system performance than
a model. Such testing would put system components in a realistic system
environment for detecting problems early during design and development. SD10 officials decided that such testing will be performed in the
System Simulator (or some other simulation framework) and will provide the basis for system design studies and end-to-end validation of
overall system performance. Simply put, the System Simulator will
increase SDIO'S confidence that when the independently designed and
developed elements become operational, they will integrate effectively
into a Strategic Defense System.
In 1988 SD10 established the Stellar Task Force to address the question
of how best to evolve to such an integrated system test environment.
While the task force fully supported the use of the System Simulator for
system-level tests, it concluded that trying to start off using only one
test environment was too complex an undertaking. It recommended
Page 22
GAO/IMTEC-90451
SD1 Needs Stable Design and Adequate
Testing
Chapter 2
Unstable Architecture
and Test Plans Will
Not Support a Presidential Deployment
Decision in 1993
breaking the problem into pieces and developing several test beds4 By
using several test beds, the system could be broken into units, each
focusing on a different part of the system. For example, a communications test bed would enable SDIOto conduct a series of tests on all of the
elements’ abilities to effectively communicate.
The task force’s rationale for this approach was that detailed test
results from each test bed could be entered into the System Simulator’s
data bases for more accurate system-level testing. The elements will be
represented at the System Simulator by software models rather than
actual element hardware and software. SDIO does plan to use several test
beds and then evolve to integrated system-level testing at the System
Simulator, incorporating actual element hardware and software in a
real-time environment. However, integrated testing will not occur before
the President’s scheduled 1993 decision.
System-Level Test Results
Based on Immature
Element Models
A stable architecture with defined system and element requirements is
required before software models can be developed to validate each of
the elements and the system as a whole. However, even prior to the
inclusion of Brilliant Pebbles, the SD10 Phase I System Engineer had
stated that detailed system requirements for all Phase I elements will
probably not be set until 1994 or 1995. Therefore, element software
models cannot sufficiently demonstrate how the various elements will
work until detailed requirements are defined. As things stand, test
results available in 1993 will be based on limited software representations of the elements. For example, Air Force officials have stated that
they have not been provided detailed ballistic missile defense requirements for the Boost Surveillance and Tracking System. Without such
requirements, system developers cannot develop an accurate software
representation for system-level testing. Consequently, it is unclear
whether detailed requirements will be determined, a model developed,
system testing completed, and results analyzed in sufficient time to support the 1993 deployment decision. Therefore, not only will system-level
tests not be conducted in real time or be integrated with pieces of hardware and software, but the tests will be based on simulations involving
immature, possibly inaccurate element models.
‘These test beds are the Surveillance and Tracking Test Bed, the Communications Network Test Red,
the Pilot Command Center. the Gaming Test Bed, and the System Simulator, all of which are
described in appendix III.
Page 23
GAO/IMTEC-90-61
SD1 Needs Stable Design and Adequate
Testing
Chapter 2
Unstable Architecture
and Test Plans Will
Not Support a Presidential
Deployment
Decision in 1993
SD10 Asserts Funding
Cuts Will Reduce
Information Needed
for an Informed
Decision in 1993
According to SDIO, an informed executive decision on deployment is contingent, in part, on minimum funding levels. At one time, this decision
was to occur in the same time frame as Defense’s formal, full-scale
development decision. In most major acquisitions, a deployment decision
is made after full-scale development is completed and the system has
moved into the next stage, (See app. I.) According to SDIO'S Director, an
informed decision on deployment is contingent on funding levels of $4.6
billion for fiscal year 1990 and $33 billion over fiscal years 1990-1994.
He stated that any funding reduction would reduce confidence in the
deployment decision and would have increasingly serious consequences
for the SD1 program. For example, the Director stated that a 20 percent
funding cut would reduce the confidence in making a decision on deployment due to cutbacks in research. He further stated that emerging concepts, especially Brilliant Pebbles, would not be fully explored. Hence,
the space-based architecture would not be completely defined, and initial system development and deployment schedules would be delayed at
least 2 years.
The Congress reduced SDIO'S fiscal year 1990 budget request by 20 percent (from $4.6 billion to $3.6 billion) because of congressional concern
for overall fiscal constraints and SDIO'S major uncertainties for fiscal
year 1990. Consequently, SDIO initiated a major replanning and restructuring strategy to identify program priorities and impacts, and SD10
delayed its full-scale development decision for 2 years. Many contracts
were reduced, delayed, or cancelled, resulting in scaled-back demonstration and validation activities that, according to SDIO, will provide less
information than was originally envisioned to support an informed 1993
deployment decision. For example, the identification of system-level
demonstration and validation requirements in support of system-level
tests will not be established until fiscal year 1992; and by 1993, testing
on command and control functions will not be as thorough as originally
planned. Further, because the new Phase I architecture will not be
defined until 1991, the elements will be less developed and their designs
less detailed, therefore, less able to support system-level testing. Finally,
the system’s communications network will not be tested for real-time
operations. Nevertheless, SDIO officials have stated that they will be able
to support a presidential deployment decision but with increased risk.
Page 24
GAO/IMTEG9@61
SD1 Needs Stable Design and Adequate
Testing
Program Management Oversight Has
ISeenReduced
SDI program decisions have been made without the formal review and
approval of the Defense Acquisition Board (DAB), increasing the risk that
program objectives will not be achieved. Defense, recognizing the complexity, uniqueness, and cost of the Strategic Defense System, required
additional DAB oversight and directed additional monitoring of SDIO'S
operational test and evaluation activities. Despite the more stringent
oversight requirements, the 1989 annual DAB review did not occur as
planned. Consequently, fundamental changes to the Phase I architecture
and reductions in test and evaluation requirements have occurred
without formal DAB review and oversight.
Oversight for Phase I
More Stringent Than
Typical Acquisition
As discussed in chapter 1, the DAB oversees major Defense acquisition
programs. As part of its oversight responsibility, the DAB conducts
formal reviews at major program milestones to ensure that programs
are ready to proceed into more advanced stages of development and
that proposed program plans for later stages follow sound acquisition
management practices, Among the issues typically considered in a DAB
milestone review are cost, requirements growth, schedule delays, threat
assessment, acquisition strategy, and test and evaluation plans and
results. As part of the Defense acquisition process, DAB reviews are to
help reduce the risks associated with major systems acquisitions.
Phase I, however, is not a typical acquisition. The complexity, cost, and
uniqueness of the Strategic Defense System has prompted the need for a
more stringent acquisition process. For example, in 1987 the DAB developed a guidance document called the Implementor which provides a
framework for additional oversight. Although formal acquisition
reviews by the DAB are usually only required at major milestones, the
Implementor expanded the DAB'S oversight responsibility to include
yearly reviews, and lists additional documentation SDIO must submit to
the DAR to ensure program goals are being met.
In addition to the increased DAB oversight, Defense also decided in 1987
to establish an oversight group under its Director for Operational Test
and Evaluation to monitor SDIO'S test and evaluation activities. Each
stage of the acquisition process, as well as each milestone decision, is to
be supported by test and evaluation The purpose of test and evaluation
is to help ensure the timely development, production, and fielding of
systems that meet the users’ requirements and perform as intended. The
DAB relies on information from test and evaluation activities to determine whether a system is ready to enter the next stage of development.
Page25
GAO/IMTEC-90-61
SD1 Needs Stable Design and Adequate
Testing
Chapter 3
Program Management
Been Reduced
1989 Annual DAB
Review Did Not Occur;
Decisions Have Been
Made Without DAB
Review and Approval
Oversight
Has
The Implementor the DAR imposed on the SD1program states that the
Phase I program baseline must be reviewed annually by the DAB. These
annual reviews were to include cost, schedule, technical performance,
and operational assessments. The annual DAB review scheduled for
October 1989 was delayed until December and later cancelled, According
to Defense officials, the review was cancelled because the fiscal year
1990 budget had not yet been finalized, and because Defense was reassessing the amount of oversight needed for the Strategic Defense System
acquisition. The review has been rescheduled for June 15, 1990. Further, significant issues have not been formally reviewed by the DAR as to
their impact on the SD1program. These issues include the October 1989
Space-Based Architecture Study’s conclusions and recommendations and
the dramatic effect of incorporating Brilliant Pebbles into Phase I.
The Space-Based Architecture Study was initiated to review the feasibility and consequences of incorporating Brilliant Pebbles into Phase I
with special emphasis on the space-based elements-the Space-Based
Interceptor, the Boost Surveillance and Tracking System, and the Space
Surveillance and Tracking System, The study concluded that Brilliant
Pebbles was feasible, but recommended that research continue on both
Brilliant Pebbles and the Space-Based Interceptor. Additionally, the
study concluded that if Brilliant Pebbles were incorporated, major
changes should be made to system requirements that could significantly
affect program cost and schedule. For example, the study concluded
that (1) the Boost Surveillance and Tracking System’s ballistic missile
defense requirements could be reduced; (2) the Space Surveillance and
Tracking System is an essential asset for Phase I but may need to be
modified; and (3) the Ground Surveillance and Tracking System’s capability will have to be increased. While SD10officials have considered the
study’s results, the DAB has not formally reviewed the study and determined its consequences on the program’s design, cost, and schedule.
IJnder Defense’s current approach, major changes to Phase I have
occurred without high-level Defense oversight and management review.
Although the study recommended that research continue on both spacebased weapons, SDIO decided to include Brilliant Pebbles in the Phase I
architecture. This fundamental change to the program baseline was
made without formal DAB review and oversight. As noted in chapter 1,
this addition will have far-reaching impacts that will ripple across the
Strategic Defense System; system interfaces, sensing capabilities, communications, and element requirements will all change significantly.
Page 26
GAO/IMTEX-!#I41
SD1 Needs Stable Design and Adequate
Testing
Chapter 3
Program Management
Recm Redur4
Oversight
Has
has also reduced the testing requirements for both the System Simulator and the Pilot Command Center without DAB review. SD10 initially
established stringent System Simulator requirements that would demonstrate an end-to-end, real-time system test with hardware-, software-,
and human-in-the loop in support of a full-scale development decision.
However, the System Simulator is now only required to demonstrate an
end-to-end simulation using software representations of the system elements instead of actual pieces of hardware and actual system software
operating in real time. Further, the Pilot Command Center, SDI’S test bed
to demonstrate battle management and command, control, and communications functions in real time, was to include actual sensor and
weapon hardware-in-the-loop testing. Current test plans do not include
actual sensor and weapon interfaces. Since these significant reductions
in demonstration and validation requirements have not been subject to
DAB review, high-level Defense Department officials have not passed
judgment on whether this proposed approach will provide sufficient
information to make an informed deployment decision.
SDIO
Congress Cuts
Funding for
Operational Testing
and Evaluation Until
Architecture Is
Solidified
The Congress has shown long-standing interest in how major weapon
systems perform and the adequacy and timeliness of operational test
and evaluation. As early as 197 1, the Congress enacted legislation
requiring Defense to provide the Congress with data on operational test
and evaluation results of major weapon systems before committing
major production dollars.
In 1983, Congress enacted additional legislation creating the Office of
the Director of Operational Test and Evaluation. Defense directed this
office to provide an independent assessment of system-level testing for
the Strategic Defense System. This office is to provide independent oversight, coordinate the military services’ planning and execution of operational tests, and objectively report on test results to Defense and
congressional decision-makers.
The Congress expects independent oversight and objective reporting of
operational test results before it is willing to commit to production decisions. However, the Congress, in November 1989, expressed concern
about the constantly changing design of the Strategic Defense System
and, given these changes, questioned what value the Office of Operational Test and Evaluation could provide to the Strategic Defense
Page 27
GAO/lMTEC-9041
SD1 Needs Stable Design and Adequate
Testing
Chapter 3
Program Management
Been
Oversight
Has
Reduced
System at that time. According to the 1990 Authorization
conferees,’
Act
...the need for a Strategic Defense Initiative operational test and evaluation activity
is premature in view of the shifting SD1architectures, plans and priorities. Accordingly, the conferees agree to reduce the operational test and evaluation authorization by $12.415 million from the requested amount and direct the Department of
Defense not to obligate any appropriated funds for such an activity until it has been
formally authorized in law.
We agree that such testing is premature given the fluid and constantly
changing nature of the Strategic Defense System architecture. However,
once the architecture is set, early system-level operational test and evaluation will help ensure that the acquisition process, which includes
system-level test plans and methodologies, is not only monitored but
also assessed. Furthermore, such oversight will help ensure that complex system-level test results, justifying a full-scale development decision, will be understood.
‘Authorizing Appropriations for Fiscal Year 1990 for Military Activities of the Department of
Defense, House of Representatives Report Iio. 101-331,10&t Congress, 1st Session (Conference
Report).
Page 28
GAO/IMTEC4081
SD1 Needs Stable Design and Adequate
Testing
Chapter 4
Conclusions and Recommendations
The Strategic Defense Initiative Organization is engaged in one of the
most complex and technologically challenging efforts ever undertaken.
Tens of billions of dollars will be needed to develop and deploy the first
phase of a Strategic Defense System. The first phase is designed to protect the United States from only a portion of a massive Soviet nuclear
ballistic missile attack.
The Phase I architecture of the Strategic Defense System remains in a
state of flux, primarily because Brilliant Pebbles, a new space-based
weapon, was added to the system. The importance of a stable architecture cannot be overstated. An architecture that is not well defined runs
the risk that system requirements will not be met and that the system
will not perform as intended. Unless the individual Strategic Defense
System elements are designed, developed, and tested as an integrated
system, SDIOwill not know whether the system will work as intended+
Brilliant Pebbles has greatly changed the Phase I architecture and highlights the instability of SDIO’S design for the Strategic Defense System.
Instead of the highly integrated, interdependent architecture envisioned
for Phase I, Brilliant Pebbles is to provide more autonomy and require
less interdependence among elements. As a result, element interfaces,
surveillance and tracking capabilities, communications, and the numbers
of elements will change. According to the Phase I System Engineer, even
before the addition of Brilliant Pebbles to the architecture, detailed
system requirements for all Phase I elements would probably not be
completely defined until 1994 or 1995. Nevertheless, Defense has
requested $265 million for full-scale development of the Boost Surveillance and Tracking System in fiscal year 1991. While the system is being
developed to meet other Defense missions, we believe that a decision to
enter full-scale development for the Boost Surveillance and Tracking
System before the Phase I architecture had been solidified and its ballistic missile defense requirements defined would be premature. Such
decisions in other Defense programs have led to significant cost
increases, schedule delays, and performance shortfalls.
Phase I system-level test and evaluation activities will be significantly
affected as well. System-level testing and evaluation are critical to
making decisions on the continued development and deployment of a
Strategic Defense System. However, the majority of system test and
evaluation efforts so far have been based on an earlier Phase I architecture that did not include Brilliant Pebbles. Consequently, much of the
current test data and analyses will not be applicable to the new architecture, and many of the tests will likely have to be redesigned and
Page 29
GAO/IMTEC-90-61
SD1 Needs Stable Design and Adequate
Testing
Chapter 4
Conclusions
and Recommendations
repeated. Even though initial SDIO test plans required that real-time integrated system-level tests be performed prior to a full-scale development
decision and a presidential deployment decision, SDIO officials now claim
such tests will not be conducted prior to either decision. We believe
SDIO’S initial plans could significantly reduce the risks involved in
making critical development and deployment decisions for such a costly
and technically complex system.
The President is scheduled to make a decision in 1993 on deploying
Phase I of the Strategic Defense System. The President will have a range
of options including deploying, delaying, or cancelling the Phase I
system. Public Law 99-145 states that the system may not be deployed
unless the President determines that the system is sufficiently capable
of fulfilling its mission. Given that the architecture continues to change,
the program is falling behind schedule, and the amount of test and evaluation data has been reduced, we do not believe that SDIO will be able to
give the President enough information to support a 1993 decision to
deploy Phase I.
We are not alone in raising concerns about the Strategic Defense
System’s changing architecture. The Congress has been troubled by the
constantly changing design of the Phase I system and has questioned the
value of operational test and evaluation oversight at this time. The Congress was so concerned that for fiscal year 1990 it not only reduced SDI
program funding by 20 percent (from $4.6 billion to $3,6 billion), but
also cut funding for the Office of Operational Testing and Evaluation,
citing such activities as premature in view of the shifting Strategic
Defense System architecture, plans, and priorities.
The federal government has repeatedly found that major acquisitions
such as the Strategic Defense System are extremely difficult and complex undertakings, fraught with high risks. Indeed, many have
floundered. A principal element needed to successfully design and
develop highly integrated systems is the commitment and oversight of
agency leaders. To provide that oversight they must have the information they need-such as program cost, requirements growth, schedule
delays, acquisition strategy, and test and evaluation plans and resultsto effectively guide the acquisition effort.
High-level oversight of the SDI program, however, is not occurring.
The Defense Acquisition Board, which oversees major Defense acquisition programs, has not reviewed major changes to the Phase I architecture and to the system’s test and evaluation requirements, The DAB’S
Page
30
EAO/IMTEC-9081
SD1 Needs Stable Desigu and Adequate
Testing
Chapter 4
Conclusions
and Recommendations
1989 annual program review was cancelled. According to Defense officials, the review was cancelled because the fiscal year 1990 budget had
not yet been finalized, and because Defense was reassessing the amount
of oversight needed for the Strategic Defense System acquisition. We do
not believe these reasons justify cancelling scheduled oversight reviews
considering the dramatic, fundamental changes in the Phase I design
and the significant reductions in system-level test and evaluation
requirements. As a result, significant changes, such as the addition of
Brilliant Pebbles to the Phase I architecture and the results of the
October 1989 Space-Based Architecture Study, have not been subject to
formal review.
Recommendations to
the Secretary of
Defense
We recommend that the Secretary of Defense advise the President to
defer a decision to deploy any element of the Phase I system until SDIO
has stabilized the architecture and has demonstrated the effectiveness
of the system through integrated system-level tests in real time, using
system hardware and software prototypes with human intervention.
Additionally, the Secretary should ensure that required oversight by the
Defense Acquisition Board be followed. The Board should more closely
monitor system design, development, testing, and evaluation. Finally,
when the program’s architecture is stable, the Secretary should request
that the Congress reinstate funding for the Director of Operational Test
and Evaluation to provide an independent system-level assessment for
the Strategic Defense System.
Recommendation to
the Congress
We recommend that the Congress not fund full-scale development for
any element of Phase I until SD10 has stabilized the architecture and has
demonstrated the effectiveness of the system through integrated
system-level tests in real time, using hardware and software prototypes
with human intervention. This would include not providing $265 million
for the Boost Surveillance and Tracking System in fiscal year 1991.
However, if Defense needs the Boost Surveillance and Tracking System
for another mission, independent and separate from the Strategic
Defense System, it should be justified and funded to meet that mission
and should no longer be considered an element of the Strategic Defense
System.
Page 31
GAO/IMTEG9OS1
SD1 Needs Stable Design and Adequate
Testing
Department of Defense Major Systems
Acquisition Stages
Mission area analysis and program initiation generally precede the five
Department of Defense acquisition stages. Defense components continually analyze their assigned mission areas to identify deficiencies (needs)
and determine if new systems or major upgrades to existing systems are
necessary. These analyses often result in recommendations to initiate
new acquisition programs through the validation of a need to correct the
deficiency. Once a need has been identified and validated and Defense
initiates an acquisition program, the program enters the concept formulation stage.
Concept Formulation
Stage
In this stage, potential requirements and alternative approaches to satisfy the need are identified and evaluated. Various types of analyses
considering trade-offs among performance, life-cycle cost, and schedule
are conducted to select among possible concepts to satisfy the need.
Once a concept has been selected, it is presented to Defense for
approval.
Demonstration and
Validation Stage
In this stage, feasibility and desirability of the selected requirements
and the system concept are further analyzed, generally using techniques
like computer simulation, hardware prototyping, development test and
evaluation, operational test and evaluation, or a combination of test
methods. When the feasibility of the concept has been convincingly
demonstrated and validated, the program enters the full-scale engineering and development stage.
Full-Scale
Development Stage
In this stage, the system, including every item necessary for its logistic
and operational support, is designed, fabricated, and tested. At the conclusion of this stage, the system is ready to be produced.
Full-Rate Production
and Initial Deployment
Stage
During this stage the proposed system is built and released to the user.
At this point, the system becomes operational.
Operations Support
Stage
This stage immediately follows deployment and extends until the
system is removed from Defense inventory. Two major Defense reviews
are conducted in this stage. The first takes place 1 to 2 years after
Page 32
GAO/IMTEC-99-61
SD1 Needs Stable Design and Adequate
Testing
Appendix I
Department
of Defense
Acquisition
Stages
Major
Systems
deployment to determine if operational readiness and support objectives
are being achieved and maintained. The second, occurring 5 to 10 years
after deployment, evaluates system capabilities and assesses whether
major upgrades are needed or if the system should be replaced.
Page 33
GAO/IMTJX-9041
SD1 Needs Stable Design and Adequate
Testing
/
i
8
D
Appendix II
Strategic Defense System Elements of Phase I
General
function
-I-...
System
element
.--..
Boost Surveillance
and Tracking
System
Space Surveillance
and Tracking
System
Ground
Surveillance and
Tracking System
Ground-Based
Radar
Space-Based
Interceptor
..--.
Ground-Based
Interceptor
Command Center
Page 34
Specific functions
..--...__
Sensor
Detect missile launches; acquire and track
boosters, assess kills
Sensor
..---Acquire and track warheads and satellites;
assess krlls
Sensor
--..
Track warheads and decoys, discriminate
warheads from decoys; assess kills
Weapon
- ..--..
._~
Acquire and track warheads and decoys;
discriminate warheads from decoys
Destroy boosters and warheads
Weapon
Destroy warheads
Sensor
-I .~
Human decision-making, communications
and guidance for defense system
GAO/IMTFX-90-61
SD1 Needs Stable Design and Adequate
Testing
Appendix III
Test Beds Supporting Demonstration and
Validation of the Strategic DefenseSystem
The Surveillance and Tracking Test Bed will evaluate tracking and discrimination function performance for accuracy, computational requirements, speed, robustness. It will also validate surveillance algorithms
and integrated system performance.
The Communications Network Test Bed will validate communications
network operations and management and validate security of the Strategic Defense System’s communications design.
The Pilot Command Center will validate command structure, hardwarein-the-loop, and demonstrate proof of concept for mobile and fixed physica1 configurations with selected software performance.
The Gaming Test Bed will evaluate proposed command and control operational concepts, build operator and user confidence in hardware-in-theloop decision tools to support the deveIopment of the Pilot Command
Center.
The System Simulator will provide the basic vehicle for system design
studies and end-to-end validation of overall system performance. It will
also validate software and integration capabilities.
Page 35
GAO/IMTEX-!W41
SD1 Needs Stable Design and Adequate
Testing
Appendix IV
Major Contributors to This Report
Information
Management and
Technology Division,
Wa+shington, D.C.
Denver Regional
Office
(510419)
Michael T. Blair, Assistant Director
Leonard J. Latham, Technical Advisor
Sally M. Obenski, Site Senior
Victoria L. Miller, Evaluator
Teresa M. Schlee, Writer-Editor
Barry A. Tidwell, Evaluator-in-Charge
Yvonne J. Rodriguez, Evaluator
Page 36
GAO/llWlECW61
SD1 Needs Stable Design and Adequate
Testing
.
.-. -_
. .
.
:
‘.:
. .
..
.:
. . .
,-. .
..
‘.
.
. .
.
:.
. .. ..
‘.’
._ .
.i
~
.
:
:
::
:
:
:
:.
. .
.
.
. . :.
-_ ..l
.- :
-
Z’..
;...
._ .
.
.
.
.
.
-
. :‘.I...”
:-_ y:-
..:
.. ..
: .::
‘._. .
-‘I.
. .
. i
United States
General Accounting
Office
Washington,
D.C. 20548
Official
Penalty
Business
for Private
Use $MO
Download